" Seismic Performance of Strength Degraded Structures"

Rupali S Bhamare
1
, Bidur Kafle
2
, Nelson Lam
3
, John Wilson
4
, Emad Gad
5
1. PhD student, Dept of Civil and Environmental Engineering, The University of Melbourne, VIC 3010, Australia.
Email: bhamarer@unimelb.edu.au

2. PhD student, Dept of Civil and Environmental Engineering, The University of Melbourne, VIC 3010, Australia.
Email: b.kafle@pgrad.unimelb.edu.au

3. Associate Professor and Reader, Dept of Civil and Environmental Engineering, The University of Melbourne, VIC
3010, Australia. Email: n.lam@civenv.unimelb.edu.au

4. Professor, Faculty of Engineering and Industrial Science, Swinburne University of Technology, Hawthorne, VIC
3122, Australia. Email: jwilson@groupwise.swin.edu.au

5. Assistant Professor, Dept of Civil and Environmental Engineering, The University of Melbourne, VIC 3010 and Fac-
ulty of Engineering and Industrial Science, Swinburne University of Technology, Hawthorne, VIC 3122, Australia.
Email: egad@groupwise.swin.edu.au


ABSTRACT:
Traditionally, performance assessment of a structure is based on trading off strength demand with
ductility demand. In high seismicity region, the design provisions are based on the concept of con-
servation of energy; such guidelines (FEMA 273) recommend a very low drift capacity for strength
degraded structures. Furthermore, the application of these guidelines, results in most of the strength
degraded structures deemed unsafe when subjected to earthquake excitations in low and moderate
seismic regions. This paper presents results of nonlinear time history analyses (THA) for such
strength degraded structures for a range of near-field and far-field earthquake scenarios of different
M-R combinations. Fragility curves defining probability of failure of structures have been devel-
oped. The insensitivity of the probability of failure of the URM wall on its height (i.e. vertical span
length) is an interesting finding in this study. Also, the extent of strength degradation in an unrein-
forced masonry wall is not shown to have increased its probability of failure.

KEYWORDS: Earthquake, displacement controlled behaviour, non linear time history analyses,
strength degraded structures.









1 INTRODUCTION

Seismic design and assessment of structures have been based on trading-off strength with ductility
demand to ensure that the structure has adequate energy dissipation capacity. In high seismic re-
gion, structures are deemed to be seismically unsafe if their energy dissipation capacity is exceeded
by the energy demand from an earthquake. Structures in high seismic regions would need to be de-
signed and detailed for ductile performances in order that they could accommodate large displace-
ments without significant lateral strength degradation. Furthermore, when such standards are ap-
plied in low and moderate seismic regions, most of the strength degraded structures would be
deemed to be seismically unsafe during earthquake excitations. For structures which have not been
designed for seismic resistance namely rigid body objects, unreinforced masonry structures (URM)
and soft-storey building, low drift capacity is recommended (ATC 40, FEMA 273) in view of their
potential strength degradation behaviour.

However in low and moderate seismic regions, velocity demand subsides as the natural period in-
creases. Importantly, the diminishing energy phenomenon indicates that the displacement demand
on a structure is insensitive to the natural period as the displacement demand is constrained to an
upper limit (Lam et al., 2005). Consequently the seismic performance of a structure can be con-
trolled by its displacement capacity as opposed to its energy dissipation capacity. According to this
phenomenon, structures are deemed seismically safe, if their displacement capacity exceeds the im-
posed displacement demand irrespective of their strength and energy dissipation capacity. Therefore
this new assessment approach based on displacement controlled phenomena should be very much
suited to low and moderate seismic regions where structures with strength degrading property are
very common.

In order to by-pass conventional force based analysis which is time consuming, the new displace-
ment controlled phenomena has been used to evaluate factors which can be used as the quick as-
sessment of structures for design purposes. The single degree of freedom (SDOF) systems repre-
senting strength degraded structures are subjected to nonlinear time history analyses (THA) using
simulated accelerograms on various soil sites. The simulations were based on a range of earthquake
scenarios defined by magnitude epicentral distance (M-R) combinations. Most importantly, earth-
quake combinations include near-field and far-field earthquakes. The main aim of this study is to
identify factors which can be applied in seismic design practices (for quick assessment of strength
degraded structures). The seismic performance of structures can be assessed simply by comparing
the displacement demand with the displacement capacity of a structure. The proposed method will
result in significant savings in time and can easily by-pass conventional force based approach for
seismic assessment. Interesting results from the parametric study have shown that probability of
failure (POF) of a wall is not very sensitive to the height of the wall whereas the POF is dependant
on the thickness of the wall and its strength degradation behaviour is shown to have not increased
the chance of overturning.

Strength degradation and nonlinear force displacement relationship of these structures are described
in Section 2. Various accelerograms with different M-R combinations employed for the analyses
are described in Section 3. The maximum displacement demand of SDOF systems can be obtained
from the elastic response spectra calculated for linear elastic systems (refer Section 4). Non linear
time history analyses have been undertaken by the author(s). Fragility curves defining POF have
been developed to study the seismic performance of strength degraded structures in displacement
controlled conditions (refer Section 5).

2 STRENGTH DEGRADED STRUCTURES
2.1 Rigid body objects (RBO)
Rigid body objects (RBO) are free standing structural assemblies. When such structures are excited
beyond the linear elastic range, the joints between the elements (or elements and the ground) will
open and leave the blocks to rock about the pivotal edges (Figure 1). Some typical examples of sys-
tems which exhibit the rocking behaviour are tombstones, monumental columns, and free standing
objects. The force-displacement relationship of rigid body rocking is generally represented by the
vertical line (a-b) which denotes infinite initial stiffness, and negative slope (b-c) showing the deg-
radation in strength as displacement increases (Figure 1). The force-displacement relationship of
rigid objects exhibits non-linear elastic behaviour as the behaviour of loading, unloading and re-
loading is similar. This indicates that objects possess excellent self-centering capabilities under
gravitational loading (Al Abadi et al., 2006).










Distribution of
inertia force

Point of incipient
rocking

Point of
instability

t

h

2/3 h

Fo

f
F

Mg
G
(a) Idealisation of rigid body objects (b) F- relationship
a
b
c

Figure 1. Force displacement relationship of rigid body objects.
2.2 Unreinforced masonry structures (URM)
Australia has a huge brick masonry building stock as compared to structures built from other con-
struction materials. However, major limitation with URM structures is their vulnerability to earth-
quake loading due to limited ductility; URM structures are also heavy and brittle with low tensile
strength. As a result, URM structures are vulnerable in areas of high seismicity. URM walls sub-
jected to one-way bending, often respond in a non-linear, fully elastic manner. These systems have
excellent self-centering capabilities as the displacement always reverts back to zero at each unload-
ing or reloading cycle. However, there is very little energy dissipation during deformation of the
walls. The force-displacement relationship is similar to that of a free-standing object, featuring sig-
nificant strength degradation and no energy dissipation (Figure 2). When subjected to horizontal
bending, a URM wall responds in a non-linear inelastic manner featuring significant strength deg-
radation but showing energy dissipation capability due to friction (Doherty et al., 2002). For analy-
ses purpose, semi rigid rocking model has been considered as it represents the real behaviour of a
structure. The parameters defining semi rigid rocking model (i.e.
2 1
, ) are calculated from the ra-
tio of
1
and
f
& ratio of
2
and
f
(Lam et al., 2003). Refer also Table 1.
Once height (h), thickness (t),
2 1
, and
f
are known, the tri-linear force displacement model
can be readily used for the analyses; where initial stiffness is controlled by
1
and strength degra-
dation is controlled by
2
. The wall defining strength degraded conditions have also been taken into
considerations. The terminology new refers to newly constructed walls, whereas the terminology
moderate and severe refers to moderately degraded and severely degraded walls respectively
(Doherty et al., 2002).

0
75
150
0 150 300 Di spl acement
F
o
r
c
e
2 1
Rigid body rocking
(bi-linear model)
Semi rigid rocking
(Tri-linear model)
Experimental
nonlinear
(real behaviour)
f
Fo
F

Figure 2. Tri linear force displacement relationship of URM.
2.3 Soft storey structures (SSC)
Building structures possessing vertical stiffness irregularity with an open plan on the ground floor
are very common. To utilise open space on the ground floor, the ground floors are free of structural
walls. These buildings behave like an inverted pendulum in an earthquake. The columns in the open
ground storey are often susceptible to severe damage, which may even lead to complete collapse of
the building. An experimental program has been undertaken at The University of Melbourne to in-
vestigate the force-displacement behaviour of soft-storey columns under cyclic loading (Rodsin et
al., 2004). Two half-scaled reinforced concrete column specimens have been tested. The span-depth
ratios of columns were deliberately considered as 3.75 (column S1) and 2.75 (column S2) so that
the columns can undergo flexural and shear behaviour, respectively (Figure 3). Both columns were
designed according to the Australian Standard (AS3600, 2001), (i.e. not designed for seismic load-
ing). The tests were terminated at the incipient collapse of columns (when the columns had lost
their axial load capacities) as opposed to the conventional 20% degradation of lateral strength (Fig-
ure 4).


Lateral
Load
Axial
Load
7
5
0

m
m

VMS
Target
R6 @ 150
200 mm
1
6
0
4Y12


Target
Lateral Load
Axial Load
5
5
0

m
m

VMS

R6 @ 150 mm
200 mm
1
6
0

m
m

4Y12



Figure 3. Configuration of column S1 and S2 (Rodsin et al., 2004) (a) Flexure dominated column (b) Shear dominated
column.
Details of the test results have been reported in the paper by Rodsin et al. (2006) which was pre-
sented in the ACMSM conference. Classical hysteretic models have been calibrated to match with
the experimental test results. The calibrated models have been used as input into the non linear time
history analyses. Program RUAUMOKO (Carr, 2003) has been used for the seismic analyses (sec-
tion 5.3). The capacities observed from the test are used to determine the state of failure for soft sto-
rey columns.
Table 1: Parameters of semi-rigid rocking model
Wall thickness
(t)=
f

= 110mm
Wall thickness
(t) =
f
= 230mm
Condition
f

1

(%)
f

1

(%)
1

(mm)
2

(mm)
1

(mm)
2

(mm)
New 6 28 7 30 14 65
Moderate 13 40 15 44 30 92
Severe 2 50 22 55 46 115
New: New condition; Moderate: Moderately degraded structures; Severe: Se-
verely degraded structures
5
5
0
m
m
7
5
0
m
m

(a) (b)

0
10
20
30
40
50
0 5 10 15 20
Displacement, mm
F
o
r
c
e
,

K
N
Prediction Experiment
0
10
20
30
40
50
60
70
0 5 10 15 20
Displacement, mm
F
o
r
c
e
,

K
N
Experiment Prediction



Figure 4. Force displacement curve from experimental test (a) Flexure-dominated Column, (b) Shear-dominated col-
umn.
3 APPLIED EXCITATION

The accelerograms for nonlinear time-history analyses (THAs) were generated by stochastic simu-
lations of seismological models using program GENQKE (Lam et al., 2000) and SHAKE (Idriss et
al., 1992). The simulations comprise various earthquake scenarios for different magnitude and epi-
central distance combinations. The combination includes far-field and near-field earthquakes. By
keeping the magnitude of the earthquake at 7 epicentral distances were varied from 177 Km to 14
Km on site class D which has a natural period of 0.89 sec (as per AS1170.4-2007 classification).
Some 30 number of M7-R combinations were incorporated into the parametric studies.

Figure 5 shows a range of magnitude-distance combinations on the same soil site. To investigate the
seismic performance of strength degraded structures, the earthquake scenarios presented in figure 5
were considered.
0
100
200
0 50 100 150 200
Epicentral distance (R) (Km)
P
G
V

(
m
m
/
s
)
M7 M6.5 6 5.5

Figure 5. Earthquake scenarios with various magnitude distance (M-R) combinations for site class D soil.
M: Magnitude of earthquake, R: Epicentral distance in Km PGV: Peak ground velocity in mm/s on rock (representing the intensity of hazard in the
area).
4 PEAK DISPLACEMNT DEMAND
Elastic systems: Simplified displacement response spectrum on rock and soil site
The maximum spectral displacement (RSD
max
) or peak displacement demand (PDD), on rock site
for low and moderate seismic regions can be estimated in accordance with the second corner period
Onset of shear
Loss axial load capacity
Loss axial load capacity
0 10 20 30 40
Displacement (mm)
0 10 20 30 40
Displacement (mm)
(a) (b)
(T
2
) and the maximum response spectral velocity (RSV
max
); where, RSV
max
can be obtained using
equation 1:
) (
max
PGV C RSV = (1)
where, C is the multiplying factor =1.8 for rock sites and PGV is peak ground velocity (Lam et al.,
2000). RSD
max
can be obtained as the intersection of T
2
and ( 2 /
max
T RSV ) (Figure 6a). where, T
2

is second corner period. RSD
max
can be considered as maximum displacement demand of linear
elastic systems on rock site by using following equation 2:
2 /
2 max max
T RSV RSD = (2)
where ) 5 ( 5 . 0 5 . 0
2
+ = M T for rock site, M is the magnitude of earthquake


T2
RSD Max
Natural Period
R
S
D

(
T
)
PGV (From Contour Maps)
RSV max =C*PGV
C=1.8 Somerville et al (1998)
C=2.0 Lam(2000a)

2
m
a
x
T
R
S
V


Natural Period (sec)
Actual
-
Idealised
Bed Rock -
Bi-linear idealisation
Tg= T2
RSD Max
R
S
D


Soil Bi-linear idealisation

(a) Simplified displacement response spectrum on rock site (b) Simplified displacement response spectrum on soil site
Figure 6. Simplified displacement response spectrum.

It has been observed from the studies that displacement demand is amplified for soil sites (Figure
6b) and the amplification factor can vary between 4 and 6 (Venkatesan, S 2006). To sum up, RSD
max on soil can simply be estimated to be amplification factor multiplied by RSD on rock site at
site natural period (T
g
).
5 SEISMIC ANALYSIS OF STRENGTH DEGRADED STRUCTURES

The single degree of freedom (SDOF) systems representing RBO, URM and SSC were subjected to
nonlinear time history analyses with hysteretic models presented in section 2 and using various
simulated accelerograms as presented in section 3. The maximum displacement demands of SDOF
systems observed from time history analyses are compared with maximum displacement capacity of
the structures. Results from the analyses were used to determine the state of failure of the structure.
The computer program RUAUMOKO (Carr, 2003) and ROWMANRY (Doherty, 2002) were used
to determine the maximum displacement demands for SSC, URM and RBO respectively. Fragility
curves which define the probability of failure of the structure are then developed to predict the drift
demand. The following paragraph describes the procedure for developing fragility curves.

Fragility curves were constructed based on the cumulative probability of exceedance function and
the log normal distribution curve fitting function. Parameters used for constructing the fragility
curves were estimated by the likelihood function (L) as expressed in equation 3.
( ) [ ] ( ) [ ]
i
i
x
i
x
N
i
i
d F d F L

=
=

1
1
1
(3)
where ( )
i
d F represents the cumulative distribution function(CDF) estimated at each RSD value.
i
x
is 1 or 0 based on the failure state that corresponds to average RSD
max
and N is the total number of
cases considered for the analysis. ( ) d F is written in the analytical form as stated in equation 4.

i
i
d
d F
ln
) (
(4)
where d represents RSD and [] represents the log normal distribution function. The CDF
evaluation is based on a unique value of and which are defined as the median and standard
deviation value respectively. These unique values can be obtained by maximising the Likelihood
function (equation 5).

0
25
50
75
100
0 100 200 300 400
Average RSDmax (mm)
P
r
o
b
a
b
i
l
i
t
y

o
f

F
a
i
l
u
r
e

(
%
)
Actual percentage of f ailure
Best fitted curve
URM: t=110mmh=3m
= 145.47, = 0.17

Figure 7. Development of fragility curve (a) determination of and using optimasation algorithm (b) fragility
curve for URM (h=3m, t=0.11m-severe condition).
0
ln ln
= =
d
L d
d
L d
(5)
The procedure described by Shinozuka et al. (2001) was used for obtaining the best fit curve. The
computation was performed by the optimasation algorithm which was implemented in MATLAB
(Figure 7a). An example of such fragility curve is shown in Figure 7b for a URM structure (h=3m,
thickness = 110mm, severe condition).
5.1 Unreinforced masonry structures (URM)
Single and double brick walls were considered for analyses; wall 1 and wall 2 were 110 mm and
230 mm in thickness. The displacement response of the structure in dynamic conditions of an earth-
quake was simulated by nonlinear time history analyses (THA). SDOF systems representing URM
walls possessed force displacement relationships as described in section 2.2. The simulation of ac-
celerograms was described in section 3. Fragility curves for aforementioned conditions were then
generated. To investigate which seismic parameter (PGV or RSD
max
) can best represent the prob-
ability of failure; wall 1 in the new condition was analysed for a range of M-R combinations for
site class D (Figure 5). It is shown in figure 8a that fragility curves for representing the probability
of failure of the walls were very inconsistent with simulations based on different earthquake magni-
tudes. Interestingly, the fragility curves became much more consistent when PGV was replaced by
RSD
max
as the parameter for charaterising the intensity of ground shaking (Figure 8b).
Secondly it is important to investigate the effects wall height has upon the probability of failure.
Walls (new and severe conditions) of different heights (2.5m, 3m, 3.5m having 110mm thick-
ness) were analysed for M 7 earthquakes with varying epicentral distances. The outcome of the pa-


(a) (b)
(a) wall 1: t=110mm
(b) wall 2: t=230mm
(a) New condition (b) Severe condition
rametric study shows that when the height varied between 2.5 m and 3.5 m, no significant change in
the fragility curves was observed (Figure 9 a,b).

0
25
50
75
100
0 50 100 150 200
PGV (mm/s)
P
r
o
b
a
b
i
l
i
t
y

o
f

f
a
i
l
u
r
e

(
%
)
M7
M6.5
M6
M5.5


0
25
50
75
100
0 50 100 150 200
AVerage RSDmax (mm)
P
r
o
b
a
b
i
l
i
t
y

o
f

f
a
i
l
u
r
e

(
%
)
M7
M6.5
M6
M5.5
0
5
10
15
20
25
0 5 10
Natur al per iod (sec)
S
p
e
c
t
r
a
l
d
is
p
la
c
e
m
e
n
t
(
m
m
)
5% damped average RSD
Average RSD max




Figure 8. Fragility curves for URM walls (wall 1 new, t=110mm, h=3000mm).
110 mm thick
0
25
50
75
100
0 50 100 150 200 250 300 350
Average RSD max (mm)
P
r
o
b
a
b
i
l
i
t
y

o
f

f
a
i
l
u
r
e

(
%
)
h=2.5,new
h=3,new
h=3.5,new
0
5
10
15
20
25
0 5 10
Natur al per iod (sec)
S
p
e
c
t
r
a
l

d
i
s
p
l
a
c
e
m
e
n
t

(
m
m
)
5% damped average RSD
Average RSD max
110 mm thi ck
0
25
50
75
100
0 50 100 150 200 250 300 350
Average RSD max (mm)
P
r
o
b
a
b
i
l
i
t
y

o
f

f
a
i
l
u
r
e

(
%
)
h=2.5,severe
h=3,severe
h=3.5,severe
0
5
10
15
20
25
0 5 10
Natur al per iod (sec)
S
p
e
c
t
r
a
l

d
i
s
p
l
a
c
e
m
e
n
t

(
m
m
)
5% damped average RSD
Average RSD max


Figure 9. Fragility curves for URM walls (wall 1: t=110mm, with h=2.5 m, 3 m, 3.5 m) for M7 earthquakes.

Wall 1 (110mm) and wall 2 (230mm) with different conditions of strength degradation and of
height 3 m have also been investigated for their seismic performance. It is observed that, the
strength degradation of the wall is shown to have not affected the probability of failure significantly
(Figure 10 a, b).
0
25
50
75
100
0 100 200 300
Average RSD max (mm)
P
r
o
b
a
b
i
l
i
t
y

o
f

f
a
i
l
u
r
e

(
%
)
h=3m,t=0.11m,New
h=3m,t=0.11m,Moderate
h=3m,t=0.11m,Severe

0
5
10
15
20
25
0 5 10
Natur al per iod (sec)
S
p
e
c
t
r
a
l

d
i
s
p
l
a
c
e
m
e
n
t

(
m
m
)
5% damped average RSD
Average RSD max

0
25
50
75
100
0 100 200 300
Average RSD max (mm)
P
r
o
b
a
b
i
l
i
t
y

o
f

f
a
i
l
u
r
e

(
%
)
h=3m,t=0.23m,New
h=3m,t=0.23m,Moderate
h=3m,t=0.23m,Severe
0
5
10
15
20
25
0 5 10
Natur al per iod (sec)
S
p
e
c
t
r
a
l

d
i
s
p
l
a
c
e
m
e
n
t

(
m
m
)
5% damped average RSD
Average RSD max


Figure 10. Fragility curves for 3 m high URM walls of different thicknesses and conditions of degradation based on M7
earthquakes.

An interesting observation to make of Figure 10 is that the extent of strength degradation in the
URM wall is shown not to have increased its probability of failure. The 5% probability of failure
limit has been considered for recommendations. The drift demand of wall 1 and wall 2 is summa-
(a) PGV as parameter (b) RSD
max
as parameter
(a) Rocking objects (b) Fragility curve

0
25
50
75
100
0 100 200 300 400
Maximum RSD (mm)
P
r
o
b
a
b
i
l
i
t
y

o
f

F
a
i
l
u
r
e

(
%
)
0
5
10
15
20
25
0 5 10
Natur al per iod (sec)
S
p
e
c
t
r
a
l
d
is
p
la
c
e
m
e
n
t
(
m
m
)
5% damped average RSD
Average RSD max
Free standing object
h=0.35m, t=0.2m
rized in Table 2. The seismic hazard factor, z (gs) shown in Table 2 is between 0.09 to 0.12 which
is comparable to the seismic hazard factor of some of the major places in Australia such as Ade-
laide, Newcastle (AS 1170.4-2007). Therefore, for 500 year return period the URM walls if prop-
erly constructed should be safe from overturning on onerous soil sites similar to the one which is
considered in this paper (site natural period = 0.9 sec). For quick assessment, it is not unreasonable
to take RSD
max
to be equal to 0.6 to 0.7 (i.e. two- thirds) of the wall thickness for design purposes.

Table 2: Seismic hazard factors for URM walls
5% probability of failure Thickness of wall (mm)
RSD
max
on soil (mm) M-R combination Z (gs) Factor
110 65-85 (M7-R85) (M7-R30) 0.09-0.12 0.6-0.8
230 125-215 (M7-R34) (M7-R21) 0.18-0.31 0.6-0.9
750 / ) / ( ) ' ( s mm PGV s g z =
,
8 . 1 /
) max(rock
RSV PGV =

5.2 Rigid body Objects (RBO)
A free standing object such as computer cabinet or library steel rack in a building may be excited
into rocking motion during an earthquake (Figure 11a). The research described in this paper is
aimed at determining the displacement of these objects and developing fragility curves. A simple
computer cabinet of 350 x 200 mm in size has been analysed for its response to ground shaking
based on M7 events. The probability of failure has been plotted in Figure 11b. The value of RSD
max
corresponding to 5% probability of failure has been identified for design purposes. It has been ob-
served from the fragility curve that this limiting value of RSD
max
was about two- thirds of the
thickness of the object (Figure 11b). The research is ongoing on RBO of varying geometry. The
POF of RBO decreases with increasing dimension (width) as reported by Al Abadi et al., 2006.









Figure 11. Free standing object subjected to M7 earthquake combinations.
5.3 Soft storey structures
Hysteretic modelling
The values of parameters defining the hysteretic models were derived by curve fitting the models to
the hysteretic loop recorded from cyclic testing. It has been found from calibration, that a reason-
able match can be observed using the modified Takeda hysteretic model. Examples of a calibrated
hysteretic model are presented in Figure 12a and 12b. The strength degradation parameters are de-
fined in Figure 12d. The parameter (r) in modified Takeda model (Figure 12c) is rate of strength
hardening which is taken as equal to zero in this study. The reasonable matches for unloading ()
and reloading parameters () of modified Takeda hysteretic model are shown as below:
For flexure dominated column: modified Takeda hysteretic model =0.5,=0.2
For shear dominated column: modified Takeda hysteretic model =0.5,=0.6
(b)
(a)
r=0
-60
-40
-20
0
20
40
60
-30 -20 -10 0 10 20 30
Di splacement (mm)
F
o
r
c
e

(
K
N
)

Cyclic test
Modified Takeda
model
Parameters:
= 0.5,=0.2
40 60 20 -40 -20
-60

-80
-60
-40
-20
0
20
40
60
80
-30 -20 -10 0 10 20 30
Displacement (mm)
F
o
r
c
e

(
k
N
)
Cyclic test
Modified Takeda
model
Parameters:
=0.5,=0.6
40 60 20 -20
-60
-40

Ku Ko
dp
Fy
+

Fy
-

r Ko
r Ko
Ku
dp
Ko
di
di

=
m
y
o u
K K or

Figure 12. Cyclic test results for soft storey columns (a) flexure dominated column (b) shear dominated column
(c) modified takeda model (d) strength degradation parameters.

The SDOF systems with hysteretic model representing soft storey column (Figure 12c) were then
subjected to non linear time history analyses using ensembles of simulated accelerograms based on
M7 events. A computer program RUAUMOKO is used for predicting the peak displacement de-
mand. Results from the analyses were correlated with average RSD
max
for the construction of fra-
gility curves which presented in Figure 13. It is shown that the drift demand of shear dominated col-
umns is higher than that of flexural dominated columns. It is not the intention of this paper to
generalize results at this stage. Research is ongoing on the potential seismic performance behaviour
of soft storey columns possessing different dimensions.

0
25
50
75
100
0 100 200 300 400
Maximum RSD (mm)
P
r
o
b
a
b
i
l
i
t
y

o
f

F
a
i
l
u
r
e

(
%
)
Flexure dominated column
0
5
10
15
20
25
0 5 10
Natur al per iod (sec)
S
p
e
c
t
r
a
l
d
is
p
la
c
e
m
e
n
t
(
m
m
)
5% damped average RSD
Average RSD max

0
25
50
75
100
0 100 200 300 400
Maximum RSD (mm)
P
r
o
b
a
b
i
l
i
t
y

o
f

F
a
i
l
u
r
e

(
%
)
shear dominated column
0
5
10
15
20
25
0 5 10
Natur al per iod (sec)
S
p
e
c
t
r
a
l
d
is
p
la
c
e
m
e
n
t
(
m
m
)
5% damped average RSD
Average RSD max

Figure 13. Fragility curves for soft storey columns subjected to M7- R combinations (a) flexural dominated column
(b) shear dominated column.
6 CLOSING REMARK

This paper presents fragility curves for estimating the probability of failure of strength degraded
components such as rigid body objects, unreinforced masonry structures and soft storey building
structures. The fragility curves were developed based on hysteretic behaviour obtained from the cy-
clic testing for soft storey columns. The observed force displacement relationships have been pre-
sented. The simulated accelerograms used for analyses were generated using stochastic simulation
considering different magnitude-distance combinations which represents both near-field and far-
field earthquakes. More than 3000 time- history analyses have been undertaken to investigate the
performance of strength degraded structures under various seismic conditions. The outcomes from
(a)
(c)
(b)
(c)
(a)
(b)
Ductility ()
RDUCT=Residual Strength as a
fraction of initial yield strength
RDUCT
1
M
u
l
t
i
p
l
i
e
r

o
n

y
i
e
l
d

f
o
r
c
e

= 4.5
= 1
(d)
the analyses reveal interesting results. The fragility curves associated with different earthquake
magnitudes were found to be inconsistent when PGV was used to characterize the intensity of
ground shaking. Parametric studies show that as the wall height varied from 2.5 to 3.5 m there was
no significant change to the fragility curve. Moreover, strength degradation of the wall is shown to
have not increased the probability of failure. For quick assessment, it is not unreasonable to take
RSD
max
to be 0.6 to 0.7 (i.e. two- thirds) of the wall thickness or thickness of the rigid body object
for design purposes. In other words the wall is deemed to be safe, as long as the RSD
max
is two-
thirds of the thickness of wall. For 500 years return period, the URM walls, if properly constructed,
should be safe from overturning on an onerous soil site similar to the one considered in this paper
(site natural period = 0.9 sec). However, the analyses have not incorporated the filtering effects of
the building. Hence, parapet walls at the roof of the buildings might be subjected to a level of haz-
ard much higher than is shown on the fragility curves presented in the paper. It is also observed
that shear dominated column has higher drift demand than flexural dominated column. Further re-
search is needed to comment on quick assessment of a soft -storey column.
7 REFERENCES

Al Abadi, H., Lam, N., Gad, E., 2006. A simple displacementbased model for predicting seismically induced overturn-
ing, Journal of earthquake Engineering, 10(6).
AS/NZS 1170.4 2007. Structural Design Actions-Part 4 Earthquake Actions, Standards Australia, Sydney.
AS 3600, 2001. Australian standards for Concrete structures, Building code of Australia, Primary referenced standard.
ATC40 1996. Seismic Evaluation and Retrofitting of Concrete Buildings, Applied Technology Council, USA.
Carr, A. J. (2003) Ruaumoko, The Maori God of Volcanoes and Earthquake, University of Canterbury, New Zealand.
Doherty K., Griffith M. C., Lam, N., Wilson, J., 2002. Displacement based seismic analysis for out of plane bend-
ing of unreinforced masonry walls. Earthquake Engineering and Structural Dynamics, Vol. 31: 833-850.
FEMA273 1997. NEHRP Guideline for the Seismic Rehabilitation of Buildings, Washington DC, Federal Emergency
Management Agency, USA.
Idriss, IM & Sun, JI 1992. Users Manual for SHAKE-91. Sponsored by National Institute of Standards and Technol-
ogy, Maryland USA and Dept. of Civil & Environmental Engineering, University of California, Davis, USA.
Lam, NTK first written: 1996 and latest edition: 1999. Program ETAMAC users manual. Department of Civil & Envi-
ronmental Engineering, University of Melbourne.
Lam, NTK first written: 1999 and latest edition: 2002. Program GENQKE users manual. Department of Civil & Envi-
ronmental Engineering, University of Melbourne.
Lam, NTK., Wilson, JL., & Hutchinson, GL 2000. Generation of synthetic earthquake accelerograms using seismologi-
cal modelling: a review. Journal of Earthquake Engineering 4(3): 321-354.
Lam, N., Griffith M., Wilson, J., and Doherty. 2003. Time-history analysis of URM walls in out of plane flexure. Engi-
neering Structures, Vol. 25:743-754.
Lam, N., & Chandler, A. 2005. Peak displacement demand of small to moderate magnitude earthquake in stable conti-
nental regions, Earthquake Engineering and Structural Dynamics, 34: 1047-1072.
Lam, N., Wilson, J., Venkatesan, S. 2005. Accelerograms for dynamic analysis under the new Australian Standard for
earthquake actions. Electronic Journal of Structural Engineering, Vol. 5: 10-35.
Rodsin, K., Lam , N., Wilson, J., Goldsworthy, H. Shear controlled ultimate behaviour of non ductile reinforced con-
crete columns, Australian Earthquake Engineering Society proceeding of the 2004 conference, Mount Gambier ,
South Australia.
Rodsin, K., Lam , N., Wilson, J., Goldsworthy, H. Collapse behaviour of columns with low aspect ratios, 19
th
Austra-
lian Biennial Conference on the Mechanics of Structures and Materials, University of Canterbury, Christchurch,
New Zealand 2006 .
Shinozuka, M., Feng, M Q., Kim, H., Uzawa T., Ueda, T. 2001. Statistical analysis of fragility curves, Technical Re-
port MCEER, Department of Civil and Environmental Engineering, University of Southern California, Task Num-
bers 106-E-7.3.5 and 106-E-7.6.
Venkatesan, S 2006. A soil amplification model for intra-plate seismic regions, PhD Thesis. Department of Civil & En-
vironmental Engineering, University of Melbourne.